For the formal power series in inverse powers , its Borel transform isTo define a Borel summation along the positive real ray, this series must first converge near zero, admit a suitable analytic continuation along that ray, and have at most exponential growth there. If , the Laplace transformconverges absolutely for . Local expansion near zero, the Gamma integral and a controlled tail then recover the original asymptotic expansion. Dividing coefficients by factorials alone does not establish these continuation and growth conditions.
For example, gives . The divergent factorial series consequently has the well-defined sum for . Scaling shows that this equals for a divergent factorial series can approximate a convergent integral; the finite geometric remainder there proves its asymptotic expansion directly.
A singularity on the chosen ray requires a contour prescription. Passing it on opposite sides can change the sum by an exponentially small residue contribution, as in Borel integral with a simple pole. That ambiguity links Borel summation to the Stokes phenomenon.
New to topics? Read the docs here!